v0.16.0

• GroqModel enum is the source of truth; constants alias it.
• Drops the deprecated curl_close calls, following @mogliabr.

v0.15.2

Documentation only. No code changes, no API changes.

The README and provider docs were recommending models that no longer exist. llama-3.3-70b-versatile was shown as the default, llama-3.1-8b-instant as the fast option, and mixtral-8x7b-32768 alongside them.

Mixtral was decommissioned on 20 March 2025, so that example had been failing for over a year. Both Llama models are decommissioned on 16 August 2026.

The docs now show MODEL_GPT_OSS_120B and MODEL_GPT_OSS_20B, which is what the provider already defaults to in code. The deprecated constants are unchanged and still shipped.

v0.15.1

Maps the neutral effort option onto this provider's own reasoning control. Until now the option was accepted here and ignored.

Additive, and a patch, so anything on ^0.15 picks it up with no constraint change.

v0.15.0

Requires papi-ai/papi-core ^0.15.

Adds the neutral effort option mapping where this provider has a reasoning knob, declares its tool-selection capability as a type so callers can ask instead of catching, and refreshes the default model where the previous one was retired.

See the commit for provider-specific detail.

v0.14.1

Adds forced tool choice, which Groq was missing since it was left out of the v0.13 release. Until now it accepted the toolChoice option and silently dropped it.

Groq's API is OpenAI-compatible, so the mapping matches deepseek and openai:

toolChoice	Sent as
'auto'	"auto"
'none'	"none"
'required'	"required"
['name' => 'x']	{"type":"function","function":{"name":"x"}}

Validation lives in core's ToolChoice::fromOption() and throws before any HTTP call. Additive and backward compatible: omitting the option sends nothing, exactly as before.

v0.14.0

Requires papi-ai/papi-core ^0.14, so this package installs alongside core 0.14. No source changes; adds the docs page.

v0.9.1

Changes since v0.9:

• Add PHPDoc to all classes and public methods

v0.9.0

Changes in this release:

• Error mapping: 401→AuthenticationException, 429→RateLimitException, 4xx/5xx→ProviderException
• Requires papi-core ^0.9

v0.8

Changes in this release

• Require papi-core ^0.8 (middleware pipeline support)

v0.7

Changes in this release

• Initial release: Groq provider for PapiAI
• Chat completions and streaming via Groq LPU
• Tool/function calling support
• Models: llama-3.3-70b-versatile, llama-3.1-8b-instant, mixtral-8x7b-32768